A Systematic Mapping Study on Business Process Variability

7/29/2019 A Systematic Mapping Study on Business Process Variability

1/21

International Journal of Computer Science & Information Technology (IJCSIT) Vol 5, No 1, February 2013

DOI : 10.5121/ijcsit.2013.5101 1

A SYSTEMATIC MAPPING STUDY ON BUSINESS

PROCESSVARIABILITY

George Valena1,

Carina Alves1, Vander Alves

2, Nan Niu

3

1Informatis Center, Federal University of Pernambuco (UFPE), Recife, Brazil

{gavs,cfa}@cin.ufpe.br2

Computer Science Department, University of Braslia (UnB),

Braslia, [email protected]

3Department of Computer Science and Engineering, Mississippi State University,

MS, [email protected]

ABSTRACT

Business Process Management aligns organisational strategy and business operation. The dynamic

environment within which organisations operate promotes changes in business processes, in a phenomenon

known as business process variability. The goal of this research is reviewing business process variability

literature to comprehend this phenomenon and analyse its theoretical foundation. Through a systematic

mapping study, 80 primary studies acted as sources of evidence to answer three research questions. By

summarizing this theoretical background, we establish a conceptual synthesis of business process

variability. We equally describe business process variability approaches and observe whether these were

empirically assessed. Finally, we discuss research opportunities in the field. Our study shows that concepts

in business process variability domain are used in an inconsistent manner, demanding a common

vocabulary. A significant number of approaches is available, but most of them lack empirical studies.

Additionally, our findings provide a diagnosis of the major challenges in the field.

KEYWORDS

Business Process Management, Business Process Modelling, Business Process Variability, Systematic

Mapping Study.

1. INTRODUCTION

Business processes have improved management activities, approximating the strategic planning

from those who execute their work to achieve organisational goals. They are the main instrumentsto organise activities and improve the understanding of their interrelationships [1]. After

introducing Business Process Management (BPM) practices, the organisation benefits from a

continuous alignment between the strategy and the implementation. This disciplined approach isgoverned by a lifecycle which models, implements, monitors and improves business processes to

reach the results desired by the institution [2].

Business processes need to be adapted as a response to evolutions in internal and externalenvironment. In this scenario, changes in business domain, new technologies or industrystandards, compliance with government regulations and stakeholders needs are examples of

change inductors [3][4][5][6][7]. The consequence of this dynamic context is typically referred to
mailto:[email protected]:[email protected]:[email protected]:[email protected]


2/21


2

as business process variability [6], which is an emergent field in BPM with many of its proposals

inspired by theories from Software Product Line (SPL) to handle process variability[8][9][10][11][12][13][14][15][16][17].

Variability in business processes is necessary in order to organisations deal with environmental

changes. However, this situation introduces challenges on both technical and business levels, andit demands enterprises to quickly adapt their processes and respective supporting systems.Managing process variability is a non-trivial task as it requires specific standards, methods andtechnologies to support process variability. These techniques increase companies

competitiveness while enabling them to keep the alignment between processes and strategies.

The work of Soffer [6], published in 2004, represents one of the first contributions to investigatethe scope of changes in business processes. Since then, a wide range of efforts has been spent on

treating the variability phenomenon. However, there has been no effort to generate a detailedanalysis of the literature in this field. Hence, we decided to investigate business process

variability through a systematic mapping study, which is a method that enables a precise review

of a broader research topic [18][19]. This analysis aimed to clarify business process variabilityand provide a detailed description of its main concepts. The central problems investigated by this

study are: how does business process variability work and what is the available support tomanage it? This statement was mapped into three research questions to define the maincharacteristics of business process variability, examine the technical support provided by BPM

literature and identify open issues in the field. Our goal is to provide relevant information aboutbusiness process variability, and allow companies and researchers to better interpret this issue.

In SPL field, variability management is addressed through modelling and execution perspectives.

Features can be statically or dynamically bound [31]. This separation is also considered inbusiness process variability approaches, with design time or runtime-oriented focuses. These are

different and independent paradigms: design-time approaches treat variations during process

definition while run-time techniques propose variability facilities during process execution. Giventhat, we decided to initially explore studies on design-time process variability.

The results of our study revealed that concepts within the business process variability field areused in a quite inconsistent and vague form. Considering that this problem impacts the

communication among academics and limits the understanding of general readers, we defined acomprehensive synthesis of business process variability notions. We also identified a significant

number of approaches for process variability management. Part of these methods is supported byautomated tools, although only a small percentage had been empirically assessed. In addition,most studies employed case studies as evaluation strategy. We also discussed the main open

issues in business process variability literature, which are centred on topics such as: automatic

verifying the soundness of business process variants, graphically representing additional

dimensions of variability in process models, introducing configuration facilities in referenceprocess models, providing decision-making support for analysts during process configuration andintroducing business process flexibility requirements in BPM tools.

The remainder of the paper is organised as follows. Section 2 briefly describes the methodologyemployed by the systematic mapping study. Section 3 presents the results, while Section 4discusses our findings. Section 5 describes threats to validity. Finally, Section 6 provides finalremarks and future works.


3/21


3

2. SYSTEMATIC MAPPING STUDY PROTOCOL

This research aimed to analyse studies in business process variability by executing a secondary

research known as systematic mapping study. According to Kitchenham et al. [19], this type ofstudy examines a broader topic and classifies the literature in that specific domain, using high

level research questions. Mapping studies main benefit is to provide research community withbaselines for further research activities while delivering value to practitioners with an overview of

a specific area. A fundamental characteristic of this type of research is basing its execution on aprotocol, states Budgen et al. [18]. Our mapping study protocol was based on guidelines providedby Kitchenham et al. [20], Budgen et al. [18], and Petersen et al. [21] and on good practices from

studies conducted by Kitchenham [22][19].

2.1. Research Questions

Our overall goal is to investigate how business process variability works and what is the supportprovided by literature. To address this objective, the following research questions were defined.

RQ1. What are the characteristics of business process variability?

The first research question (RQ1) aims at offering a theoretical understanding of business processvariability phenomenon. By describing business process variability and the notions it

encompasses we aim to support the definition of a common language to improve communicationamong researchers and practitioners.

RQ2. What are the available approaches for business process variability management?

This research question evaluates the theoretical and practical support offered for practitioners todeal with business process variability. It is dismembered in three sub-questions:

RQ2.1. What are the characteristics of the approaches?

This sub-question strives to describe the approaches for treating business process variability.Additionally, the classification of the proposals according to their main focus helps practitionersto assess the available support [33] concerning a particular variability approach.

RQ2.2. Is automated tool support available?

The aim of this sub-question is to investigate the availability of automated tool support provided

by the variability approaches. It lists applications which are freely offered and also those

presented as proprietary tools or as plug-ins for proprietary BPM suites. This may supportorganisations to evaluate technologies [34] for business process variability management.

RQ2.3. How are the approaches empirically evaluated?

This sub-questions goal is revealing the percentage of studies that were empirically assessed.Highlighting proposals with an empirical evaluation and the methods used in their assessment can

play an important role in transferring research outcomes into practice.


4/21


4

RQ3. What are the current challenges within business process variability field?

The goal of RQ3 is to explore the current challenges in business process variability topic. It

strives to describe research opportunities regarding variability management and invite researchcommunity to address them. It complements the overview of the state-of-the-art in RQ2.

2.2. Search Process

As proposed by Kitchenham [20], searches for primary studies generally start employingelectronic databases. To improve the quality of search results, we have also manually analysed

the main venues for BPM researchers to publish their results. These acted as additional sources ofevidence. Both automatic and manual searches were not restricted by publication year. The searchprocess is presented in Figure 1, being described as follows.

Figure 1. Search process

Step 1 Automatic Search

During this step, we searched the following databases: ACM, Wiley InterScience, SpringerLink,ScienceDirect, CiteSeerX and IEEEXplore. The search string used in this procedure was

composed of two parts: process AND variability. To improve this initial structure, we determinedsynonyms, related or complementary terms and alternative spellings. These additional keywordswere incorporated using OR and provided a more complete query. We have tested the new search

string through a sanity-check, introducing the expression in part of the databases to assure itsstructure worked as expected. The final search string is presented below.

The automatic search returned 13.619 papers, with the following division: 30% obtained from

ScienceDirect, 23.3% from IEEE, 16.2% from Wiley InterScience, 11% from Springer, 10.4%from ACM and 9.1% from CiteSeer. An initial manual filtering was then executed to refine theresults. The criteria applied in this procedure considered papers type, title and venue in which it

was published. First, the type of the reference was verified: only papers published on conferences,

journals or workshops had their titles evaluated. Next, since this review only included studiesinvestigating business process variability in BPM or Information Technology fields, we

considered important to analyse the source where each study was published. Hence, papersbelonging to out of scope venues (e.g. Climate Dynamics) were discarded. As a result of this

procedure, 6.8% (928) of the papers were retained. Since in the resultant list we found papersfrom different venues reporting the same study, an initial application of the second inclusion

criterion (see Section 2.3) was necessary. Accordingly, 623 studies were retained.


5/21


5

Step 2 Manual search

We established a list of 13 journals and 6 conferences for the manual search, over which the

manual search was performed (Table 1). This procedure retrieved 146 papers, which had as mainsources: Intl Conference on BPM (37.7%), Intl Conference on Advanced Information Systems

Engineering (25%), Intl Conference on Conceptual Modeling (5.5%), Data and KnowledgeEngineering (4.8%) and Intl Journal of Cooperative Information Systems (4.8%).

Step 3 Merging of studies lists

This step aimed at merging the automatic and manual search results and discard duplicate papers.

We aggregated 623 papers from the automatic search and 146 from the manual search,with a resultant list of 711 papers (the absence of 58 items was due to overlaps in input lists).

Table 1. Searched journals and conferences.

Type Venue

J Business Process Management Journal

J Communications of the ACMJ Data and Knowledge Engineering

J IEEE Software

J IEEE Transactions on Knowledge and Data Engineering

J IEEE Transactions on Software Engineering

J IEICE Transactions on Information and Systems

J Information and Software Technology

J Information Systems and e-Business Management

C Intl Conference on Advanced Information Systems Engineering

C Intl Conference on BPM

C Intl Conference on Conceptual Modeling

C Intl Conference on Software Engineering (ICSE)J Intl Journal of Cooperative Information Systems

J Intl Requirements Engineering Conference (RE)

J Intl Software Product Line Conference (SPLC)

J Journal of Systems and Software

J Requirements Engineering Journal

J Software and System Modeling

Step 4 Additional studies inclusion

The goal of this step was to include papers that were not mapped by search databases or which

were not identified during manual searches (e.g. PhD and master theses addressing business

process variability). Therefore, we included 37 additional works identified on references ofprocess variability papers obtained in steps 1 and 2. A fourth list with 748 papers was generated.

Step 5 Studies selection

During this step, we applied inclusion and exclusion criteria based on the abstract, introduction

and conclusion of each paper. From the 748 papers, 127 were retained for further analysis. Toretain studies exploring design-time process variability and discard those exclusively focused on


6/21


6

the runtime environment, we applied a filter on the abstract, introduction and conclusion of the

papers to analyse their scope. We also consulted some authors to confirm our findings and discussthis separation. Accordingly, we obtained 80 primary studies. The complete list of selected papers

is described in Appendix A.

2.3. Selection Criteria

The criteria list aims at supporting the selection of the primary studies to be analysed in themapping study. We present below the set of criteria applied.

Inclusion Criteria (1) The study approaches the variability issue and/or one of the topics in the

search string; (2) Where several papers reported the same study, only the most complete oneshould be included; (3) Where several studies were reported in the same paper, each relevantstudy was treated separately; (4) Studies that answer at least one research question.

Exclusion Criteria (1) Studies that are not written in English; (2) The paper is outside the

business processes field; (3) Whitepapers, books, posters, summaries of articles, tutorials, panels,presentations, personal opinion pieces and/or viewpoints were excluded; (4) Any study not

accessible or not available in PDF or Microsoft Word format.

2.4. Data Extraction and Analysis

After the search and selection procedures, the primary studies were examined through an

extraction form. The papers were analysed considering the information required by each researchquestion. The extraction spreadsheet was filled with text excerpts from the primary studies toanswer each question. We have extracted the following data from each study: (1) Source, title,

authors and year of publication; (2) Business process variability definition, related ideas andrelationships; (3) Name and brief description of the proposed approach; (4) Availability of a

software tool. In case there exists tool support, we further identify whether the tool is open-

source or proprietary; (5) Empirical methods employed to evaluate the approach; (6) Open issueswithin business process variability topic.

Individual textual files were generated to assemble the information regarding each research

question. To analyse data, an open coding procedure was conducted [16]. This strategy has thepurpose of generating categories through the division of data gathered, yielding several concepts.

These definitions can then be modified, given comparisons and merges of notions.

3. RESULTS

3.1. Overview of Studies

No time restrictions were defined for the papers to be gathered by the search process. However,

the final list of papers was limited to the past decade, with studies from 2004 until 2011. The

investigation of business process variability was in continuous rise from 2004 to 2008. In the last3 years the number of studies published in this topic experienced a slight decrease.

Concerning the source, among the 80 primary studies, 37.5% were published in conferenceproceedings, 30% were originated from workshops, 20% were published in journals, 8.75% weretechnical reports and 3.75% were academic theses. The predominance of conference papers and

the large number of workshop studies suggest that research on business process variability is at aninitial stage. In addition, the 80 primary studies were associated to 142 different authors. Among


7/21


7

these, it was possible to identify that Marcello la Rosa, Marlon Dumas, Selmin Nurcan and

Wil M.P. van der Aalst were the top contributors to business process variability field.

3.2. Evaluation of Research Questions

We discuss in this section the results from the assessment of the primary studies, which wereobtained by applying the protocol detailed in Section 2. This analysis considered the information

required by each research question.

3.2.1. RQ1: What are the characteristics of business process variability?

This research question aimed to explore business process variability as a concept and provide aclear definition of this phenomenon. Passages from the primary studies that provide a theoretical

background on business process variability were integrated and evaluated in an open codingprocedure. Theme categories represented the available knowledge with respect to business

process variability and summarised concepts and theories within this topic. Drawing theseconclusions, we may direct efforts of BPM researchers towards a more conscious and coordinateduse of conceptual knowledge.

Business process variability is the capability of an artefact to be configured, customised orchanged for use in a specific domain. Given its conceptual foundation on reuse-oriented

development in BPM, it enables the reuse of parts of a model while adaptations to its functioningare introduced. Many authors use the term flexibility to refer to the notion of variability in thescope of business processes. Similar to business process variability, business process flexibility

is the ability of a process to adapt to the changes in the environment or to its changingrequirements. It concerns how rapid and easy a process model is modified.

Several authors split business process variability in two perspectives: design-time and runtime.The former type essentially refers to variations of models during modelling phase, before they are

implemented in a workflow management system or BPMS for execution. The latter type isassociated with processes on execution, addressing runtime variability with exception handling

approaches. These types are complementary to each other and can also be called design-time andruntime business process flexibility, given the tiny boundary between variability and flexibility

notions.

A reference business process model supports the reuse paradigm by collecting and depictingproven best practices of a specific domain. It provides a starting point to define process models

for a particular setting (e.g. a company) and improves modelling by avoiding the construction of amodel from scratch. However, the common traits captured by these models do not turn them into

plug and play solutions. Adjustments must be executed, since these generally do not offerconfiguration facilities.

The concept of configurable business process model implements the notion of reference

business process models. It is a step forward towards the reuse of business processes. They aredefined via a configurable process modelling language or notation, which provides means to

insert variability in a process model. Configurable models are constructed by merging severalbusiness process variants, which are processes achieving the same goals but slightly differingfrom each other in their structure due to domain specific requirements. Variants are versions of a

particular process model called business process type, which is defined at design-time andrepresents a standard way of acting within an organization. Variability is achieved by introducing

placeholders in a configurable model referred to as variation points, also known as configurable


8/21


8

nodes or adjustment points. Options are assigned to them by means ofvariability mechanisms,

which are techniques (e.g. extension) that realise variability in a model.

To customise configurable models to a specific solution, an analyst selects the most suitable

option for each variation point, in a procedure called business process configuration. It derives a

process from a configurable model by restricting its behaviour and can be regarded to as design-time business process variability. The scheme spawned via a configurable model is a process

variant, which is enacted as a business process instance at runtime. If changes are implementedduring the execution of this process instance, this is referred to as runtime business process

variability.

Process configuration implies decision making and it is therefore guided by configuration

decisions. These judgements are applied over each variation point to assess its available choices,based on information from the context in which the derived model should be employed. This

information may be expressed as configuration requirements (hard constraints) andconfiguration guidelines (recommendations), which can be bound to variation points to restrictthe combination of available options.

Configuration requirements and guidelines aim to avoid undesirable configurations. This leads tothe notion of business process correctness, which is a characteristic of the generated processmodels being valid in a syntactic and/or semantic form. The syntactic property is centred on theadequate use of the modelling notation through which the model structure was created. The

semantic property is also called business process soundness, analysing the dynamic behaviour of

the process model to ensure that no deadlocks or livelocks in the control-flow prevent a proper

completion.

The collection of variants obtained by means of process configuration can be denoted as business

process line. These models represent alternative forms of the same underlying process and sharean invariant nucleus known as core process. This common structure expresses the compromiseskept by members of a process line. In addition, the degree of commonality a process model keeps

with respect to another model within a process line is known as business process similarity. It

results from the comparison of multiple aspects of process variants to describe to which degreethey share a similar structure.

3.2.2. RQ2: What are the available approaches for business process variability

management?

The results of this research question provide an overview of the proposed solutions to handlebusiness process variability. We identified a set of approaches and supporting tools in the selectedstudies, which are classified and briefly described below. Additionally, we analysed the empirical

evaluations conducted by business process variability studies in order to assess the rigour and

feasibility of the proposed approaches.

RQ2.1: What are the characteristics of the approaches?

Fifty-seven (71%) among the 80 primary studies developed an approach for business processvariability management. Less than a third of the selected studies (23) did not provide a specific

approach. These studies proposed surveys (e.g. survey of flexibility requirements (P6)) or generalevaluations (e.g. assessment of techniques for evaluating process similarity (P9)), investigations(e.g. analysis of process change scope (P11)) and formal discussions (e.g. Weicks theory (P13)).

However, it is relevant to remark that studies proposing novel approaches also brought abackground concerning theories behind the practical perspective.


9/21


9

This review identified 57 unique approaches to deal with business process variability from

different aspects (e.g. process configuration, process flexibility, etc.). Although the studiesaddressed different issues, they rarely gave the same emphasis to these perspectives. Hence, to

conduct a better analysis, the approaches were classified considering their main focus (in bold),which follows the aspects of business process variability presented in former section

.A set of approaches investigated business process configuration procedure, which aims toobtain a variant from a process reference model or configurable process model. To support this

activity, some studies also explored domain aspects and decision-making aspects. We includedthe following 10 studies (around 18%) in this category: P19, P26, P27, P34, P36, P39, P45, P47,P50 and P61. These works analysed how to assist the individualization of a process model to a

particular context. In order to enhance this activity, the use of non-functional requirements wasproposed as constraints that the process must comply with during instantiation (P27). It was also

found a decision support model in the form of a questionnaire-driven approach independent ofprocess modelling notation (P68). Another study proposed a description of change patterns to be

applied during model customization (P19). The use of views by means of queries (P61) and the

control of context factors through the use of decision tables (P39) were also examples of solutionsproposed in the selected studies.

Business process correctness aims to guarantee that process models are syntactically andsemantically correct. With the lowest number of contributions, this topic encompassed 6 studies

(around 11%): P29, P41, P60, P62, P68 and P79. One of the techniques in this group was aframework whose goal was not to check the correctness of a single business process model, but to

ensure the soundness and semantic validity of a group of process variants (P41). Anotherapproach used configurable versions of EPC models (C-EPC) and supported their adaptation via a

mapping to a lawful regular EPC (P60). Additionally, two other approaches focused onconfiguring reference process models in a correctness-preserving manner (P62, P79).

An increase in flexibility is achieved when a process can be changed in a fast and easy form. Tenstudies (almost 18%) were identified with proposals focused on treating business process

flexibility: P4, P5, P7, P8, P20, P21, P30, P44, P58 and P74. Some approaches employed

business rules concepts (P5, P7, P20), with one of these splitting process behaviour into a stableand a flexible part (P74). An algebraic framework based on Algebra of System (AoS) was used to

achieve the flexibility in business processes, with process models decomposed into differentmodules of knowledge and encoded as different algebraic domains (P8). In addition, a task

mining technique was applied on a process model by reverse engineering it in a Petri net (P44).Business process variability modelling studies investigated the graphical representation of

variability, generally exploring a process modelling notation (e.g. EPC) or adapting a language

from a different field (e.g. Feature Models). Twenty-four approaches (42%) were included in this

group: P3, P10, P12, P14-P17, P22, P23, P25, P37, P40, P43, P46, P48, P49, P51, P54, P55-P57,P63, P69 and P77. Examples of these contributions are the introduction of features in EPCmodels (P3, P22, P37, P57), a hierarchical representation method for UML 2.0 activity diagrams

(P56) and, extensions of BPMN (P12, P49, P77) and MAP models (P14) with constructs

supporting variability modelling. Some approaches emphasised the importance of notationindependence, such as a framework which manages variability and supported process modelsreuse (P50).

Approaches addressing business process similarity shared the goal of diagnosing commonalitiesand variations among business process models. Seven studies (12%) were included in this group:

P33, P38, P42, P52, P53, P67 and P72. One of the proposals focused on quantifying the

similarity based on the sequence of activities holding for the process model (P52). Another


10/21


10

technique classified processes as relevant, irrelevant or potentially relevant to a search

query model (P38). Additionally, to detect similarities between process models, an approachused comparisons of the linguistic structure of their elements (P42).

RQ2.2: Is automated tool support available?

This research question examined the availability of automated tool support among the selectedstudies. Less than half of the approaches (45.6%, 26) provided a mechanism to automatically

support their proposals: P4, P10, P16, P17, P19, P22, P23, P34, P36, P38, P40, P41, P42, P47,P48, P49, P50, P52, P53, P54, P57, P44, P60, P61, P62 and P68.

Sixteen (61.54%) among 26 tools are free and open-source (P4, P16, P22, P23, P34, P38, P40,P42, P49, P50, P52, P53, P57, P61, P62, P68). These tools can be downloaded from dedicated

websites, with manuals and tutorials. Additionally, among proprietary proposals (P10, P17, P19,P36, P41, P47, P48, P54, P44, P60), some were conceived as extensions or plug-ins for popular

BPM suites, such as ARIS (P19, P36, P41, P60). By free offering the tools or improving popular

BPMS researchers promote a wider adoption of their solutions by practitioners, since these can beapplied in real scenarios.

RQ2.3: How are the approaches for business process variability empirically evaluated?

This sub-question focused on the number of approaches empirically analysed. It describes themethods used for assessing the solutions presented by the 57 studies identified in RQ2.1. Only 17

(30%) of the solutions were experimented in practice: P15, P16, P17, P19, P22, P23, P33, P37,P38, P39, P46, P48, P50, P52, P54, P68 and P79.

Three different strategies were applied in the empirical evaluations, individually or in a combined

form: case studies, experiments and surveys. Case study was the most popular method, being

used to assess 14 proposals: P15, P16, P17, P19, P22, P23, P33, P39, P46, P48, P50, P54, P68and P79. The industrial context was the preferred setting for carrying out case studies, with

approaches being tested in automotive (P19 and P54), film (P22, P51 and P68) and medical (P54)

domains, for instance. There are also case studies executed in a government setting (P23) andbased on data obtained from a literature analysis (P46). Two studies used a multiple case study

design (P19 and P54). In comparison with single data source studies, this allows an improvedjustification of findings while generating additional results, as remarked by Bratthall and

Jrgensen [24].

Three studies (P16, P37 and P38) used an experimental design. Two of them executed multiple

experiments, which is best thought of as replications. This increases the credibility of the study,

allowing more robust conclusions to be drawn [25]. In addition, surveys were used by only two

studies (P23 and P52), with different data collection techniques. A focus group was carried in onestudy (P23) as a means to test the practical usefulness of the approach, where the results werediscussed with software providers and consultants. Another research (P52) distributed an on-line

questionnaire among process modellers.

It is important to note the existence of studies using a mixed methods strategy: P16 proposed aresearch design with an experiment followed by a case study, while P23 conducted a case studyand further evaluated additional aspects of the solution using a survey. In both cases, data

collection and analysis were supported by quantitative and qualitative method. As discussed byEasterbrook et al. [23], this research strategy emerged in the recognition that all methods have

deficiencies, and weaknesses of one empirical technique can be compensated for by the strengths

of other techniques.


11/21


11

The lack of empirical analysis in 70% of the solutions was balanced by the provision of

examples. All studies proposing a novel approach presented examples illustrating its operation.These samples were classified in two groups: short examples (45.6%) and working examples

(54.4%). The first category represents studies with brief descriptions of the approach operation,eventually contextualized with figures. On the other side, studies with working examples offered

a detailed explanation of the method and simulated a real case with a step by step demonstrationof the approach.

3.3. RQ3: What are the current challenges within business process variability field?

Applying this research question to the set of primary studies enabled us to comprehend the

current direction within business process variability field and provide a diagnosis of its majorissues. These problems represent challenges raised by academics and potential research

opportunities. Based on inputs from P1, P6, P19, P20, P23, P27, P28, P32, P34, P36, P39, P41,P45, P46, P47, P50, P62, P73 and P77, we depict these challenges as follows.

Guarantee the correct configuration of a whole process family is a challenge which hasreceived little attention. This implies on a difficult and time-consuming task: to ensure the

correctness of all processes obtained from a particular configurable model (P41). A naiveapproach to treat this issue would require solving an exponential number of state-space problems(P34). Also, hiding and blocking mechanisms (P2, P24) applied on fragments of a process model

may promote behavioural anomalies such as deadlocks and livelocks. This is exacerbated by thenumber of possibilities to configure a process model, the complex domain and data dependencies

between configuration options (P34). Additionally, with manual methods for processconfiguration, analysts are left with the burden of ensuring the correctness of the customized

models and of manually fixing errors (P50).

Business process flexibility requirements need to be addressed in order to decrease reactiontime for process change. To ensure flexible business process support one must deal with amyriad of business requirements (P20). In addition, the identification, documentation and analysis

of flexibility requirements are not trivial activities (P28). To treat these aspects, we identified a

structure summarizing flexibility requirements to be handled by process-centred systems (P6).Also, it was addressed how the need for flexibility affects business process flexibility

requirements (P32). These are relevant contributions to obtain flexibility in an automatedsetting. To achieve that, process design and realization mechanisms are required to preserve

business invariants during changes (P73) and provide guidance for stakeholders to define flexiblebusiness processes (P1).

Another challenge is to develop modelling techniques that consider the stimulus for changeand not only capture the reactive part of flexibility . This involves recognizing context changes

together with knowledge about which types of change lead to an increased flexibility anddecreased reaction time (P28).

A major issue for reference modelling is the lack of sophisticated concepts and tools to support

process reuse. Analysts must spend huge amounts of time on adjusting a standard solution to theindividual needs of the organization (P23). Accordingly, the absence of an explicit representationof configuration possibilities and decisions generates an entirely manual individualization, withreference models merely as a source of inspiration (P62).

In the context of process similarity, a general concern is comparing templates to decide on

which process variant to implement. Given the lack of an explicit representation of differences


12/21


12

between similarly structured templates, manually finding the small differences can be a difficult

and time-consuming task (P23).

There is a lack of adaptability in formalisms. Since business processes provide an integratedview over an organisation, modelling languages must enable the definition and selection of

several aspects (e.g. tasks, resources) (P77). On the opposite direction, excessively describingbusiness aspects within the configuration layer makes the instantiation of a model with a largenumber of variation points close to unmanageable (P50). Hence, a proper level of abstraction

must be defined to minimize model complexity and interpretation problems.

A related challenge is choosing between specifying process variants within a single model and

defining them in individual artefacts. In most cases neither the use of separate models forcapturing different variants nor the description of variants using conditional branches constitutes

a viable solution (P19). The first option results in a huge amount of redundant and unrelatedmodel data. This complicates variant management (P45) and demands BPM tools to support

variants modelling and maintenance (P36, P39, P41). The definition of all process variants in a

single structure makes variants hard-wired in the control flow logic. This strategy also generates alarge model for a particular process family, which is difficult to comprehend and expensive to

maintain (P19). No comprehensive solution is available to adequately model multiple processvariants within a unique structure (P47). A potential solution to both cases could be theconfiguration of process models using explicit change operations and adjustment points (P19).

The adaptation of process models due to changing conditions has to be quick and precise (P2,

P12, P67, P76). The usual complex and continual transformations demand great ability fromanalysts. Frequently, analysts have to adapt process models while considering risks and costs

involved in process configuration (P46). To address this challenging dynamic context, it isimportant to offer guidance and decision support during configuration activities. Despite the

relevance of this theme, process configuration is generally performed in an ad-hoc basis

guided by the analyst experience. Although some techniques have been proposed, such as theusage of questionnaires and domain analysis, these are mainly concerned with the elicitation of

variability than the configuration procedure (P27). The lack of explicit links between variation

points in the model and business needs turns difficult to estimate the impact of configurationdecisions. The user must then possess expertise in the application domain and in the modelling

notation (P50).

4. DISCUSSION

We provided in this paper an overview of relevant concepts for business process variability bysummarizing theoretical background accumulated in the domain. This synthesis provides a

catalogue of notions composing variability context and clarifies the connections among them.This analysis revealed that definitions on process variability were scattered at diverse studies,

being employed in an inaccurate and redundant way. The concepts described in this initialtaxonomy can produce a standard vocabulary. In this sense, it can be used to improve thedialogue among practitioners (i.e. communication between process analysts, designers,

developers and users) and academics. Similar efforts have been conducted in studies P64, P65and P75, which focused on process flexibility context to define patterns and taxonomies.

We found that a large majority of studies (71%, 57) proposed solutions for process variability

management. Their main focus was on representing variability within process models (42%, 24).

To treat this issue, most approaches proposed extensions to popular process modelling notationssuch as BPMN and EPC. Complementing this aspect, solutions supporting the configuration

procedure and flexibility characteristics appeared as a second trend (18%, 10 each). Some


13/21


13

proposals for business process configuration emphasised the relevance of decision-making

support by using questionnaires and decision tables. Among solutions addressing businessprocess flexibility we perceived a predominance of modelling formalisms such as Algebra of

Systems, Petri nets and business rules. Finally, business process correctness and similarity wereaddressed by only a few studies (6 and 7 proposals, respectively). Despite the low number of

contributions, the proposed approaches were not in early stages of development. We identified,for instance, mature frameworks for ensuring well-formedness of a group of process variants andtechniques to semantically evaluate similarities between process models.

Development efforts are centred on manipulating the variability with the insertion of variationpoints in process models. Process correctness and similarity have currently a potential lower

priority in regards to variability representation, process configuration and flexibility aspects. Aprobable reason for this result is the origin of basic constructs from several approaches for

process variability management in SPL field. While variability modelling and related concernsare consolidated in SPL approaches, the definition of well-formed product lines is a topic which

is starting to be explored by the SPL community. The extent to which this issue is explored is still

mostly type safety, i.e., syntactic. In SPL approaches, semantic issues are only beginning to betackled automatically by the use of model checkers [26].

Almost 47% (26) of the 57 studies providing approaches for business process variability offeredautomated tool support. We argue that approaches are more powerful if they are supported by a

software tool, since it offers guidance to process designers in handling business processvariability. Otherwise, the use of the proposed methods tends to be time and resource consuming.

Hence, it is more suitable to exploit automated tools as a means to obtain high quality processmodels, enable designers to evaluate the effects of the proposed solutions and, finally, use these

tools as learning support for adopting the approaches. Positive aspects in the offered tools are thatmost of them were freely available (61.5%, 16) and some of them were conceived as extensions

of popular BPM tools (15%, 4), which helps to broaden the dissemination of the correspondent

methods.

Concerning empirical evidence of the approaches, we observed a low number of evaluations

(32%). This can be viewed as a considerable problem, since professionals often need empiricalevidence to judge if the technique is suitable to their context [32]. Hence, the absence of

evidence regarding practical effects of the approaches limits their acceptance. For researchers,these findings are relevant for replicating the evaluations and can serve as input to further

investigate the applicability of the techniques with different strategies.

Our findings revealed a preponderance of qualitative research methods, which were employed in

84% (14) of the empirical assessments. This reflects the recent and growing adoption of

qualitative methods in Software Engineering, Human-Computer Interaction and Information

Systems fields, as stated by Seaman in [27]. A main advantage of qualitative research isproviding researchers with richer and more informative data.

The qualitative strategies were composed by case studies and surveys, with a preponderance of

the former (74%). Although case studies cannot reach the scientific rigor of formal experiments,they can provide practitioners and researchers with sufficient information about the benefitsoffered by a specific technology to an organization or project, according to Kitchenham et al.[28]. Experiments were conducted by only a few studies (16%, 3). This might be due to the

difficulty to execute formal experiments when the degree of control is limited, which leads tosmall experiments in a real scenario. On the other hand, case studies avoid scale-up problems in

industrial evaluations, which happen when changing from a laboratory to a real context [28].


14/21


14

We also noted the use of mixed methods in empirical assessments. Among the 3 experiments

conducted (16%), one of them was followed by a case study, adopting a mixed method approach.This highlights the relevance of qualitative studies to answer questions that involve variables

which are difficult to quantify (e.g. why questions) and which have already being addressed byquantitative research [27]. In addition, one study executed the empirical evaluation by combining

case study and survey strategies. Surveys combine advantages of experiments with those of casestudies, such as replication that minimises the unusual results [28].

Our mapping study also investigated practical aspects of business process variability as a meansto detail issues not yet addressed in the literature. We observed a clear demand for automaticsupport to verify the soundness of a business process variant obtained by process configuration.

Also, reference models must be equipped with configuration facilities to act not as a single sourceof inspiration but rather as a configuration instrument. The variability represented in process

models should involve additional aspects (e.g. organisational resources) while balancingexpressiveness and complexity. Additionally, process configuration must be supported by

mechanisms providing more guidance and BPM systems should handle requirements for process

flexibility. We aimed to foster insights of future works in academic projects by highlighting theseresearch gaps. These novel studies may trigger the improvement of existing approaches with a

view to provide a greater support for business process variability.

5. THREATS TO VALIDITY

A potential risk is associated to the selection step of the search process. Although based on apredefined set of criteria, this stage was guided by the experience of the research team. Given

that, relevant studies may have been missed, notwithstanding our aim to ensure the completeness

of the selection. Additionally, during data extraction procedure, to answer the research questionsit was necessary to interpret the subjective information provided by the studies in several

occasions. This happened because many studies did not present objective details regarding theissues investigated. In order to prevent inaccuracies and minimize a potential bias, we discussed

extracted data to reach a consensus. Additionally, the conclusions obtained by a given researcherwere further evaluated by at least another member of the team.

Another potential threat to validity concerns the electronic databases employed in automatic

search procedure. Natural limitations of search engines may have caused the loss of relevantpapers. To mitigate this issue, a manual search was conducted to improve the quality of search

results. Whitepapers, posters, summaries of articles, tutorials, panels, presentations and personal

opinion papers were excluded from this review. Discarding this literature had probably no impacton research results, as this is unlikely to provide relevant and/or mature information.

6. CONCLUSIONS

Business process variability can be viewed as a result from the dynamism of the organisational

and business environment. It is the ability of a process model to be adapted for use in a specificdomain, employing existing knowledge and reducing response time and modelling efforts [29].

Variability management within BPM has been studied since the last decade, with a significant

number of approaches reported. This paper aimed to investigate business process variabilitythrough a systematic mapping study.


15/21


16/21


16

[13] La Rosa, M., van der Aalst, W., Dumas, M., ter Hofstede, A. (2009) Questionnaire -Based Variability

Modeling for System Configuration, Soft. and Systems Modeling 8, pp. 251-274.

[14] Rolland, C., Nurcan, S. (2010) Business Process Lines to Deal with the Variability, Proc. of

HICSS, pp. 1-10.

[15] Boffoli, N., Caivano, D., Castelluccia, D., Maggi, F. M., Visaggio, G. (2008) Business Process Lines

to Develop Service-Oriented Architectures Through the Software Product Lines Paradigm, Proc. of

Software Product Line Conference, pp. 143-147.[16] Hallerbach, A., Bauer, T., Reichert, M. Capturing Variability in Business Process Models: the

Provop Approach, Journal of Software Maintenance 22 (6-7), pp. 519-546.

[17] Kulkarni, V., Bara, S. (2010) Business Process Families Using Model-Driven Techniques, Proc. of

Business Process Management Workshops, pp. 314-320.

[18] Budgen, D., Turner, M., Brereton, P., Kitchenham, B. (2008) Using Mapping Studies in Software

Engineering, Proc. of PPIG, pp. 195204.

[19] Kitchenham, B., Budgen, D., Brereton, O. (2011) Using Mapping Studies as the Basis for Further

Research: A Participant-observer Case Study, Information and Soft. Technology 53 (6), pp. 638

651.

[20] Kitchenham, B., Dyba, T., Jorgensen, M. (2004) Evidence -based Software Engineering, Proc. of

the Intl Conference on Software Engineering, pp. 273 -281.

[21] Petersen, K., Feldt, R., Mujtaba, S., Mattsson, M. (2008) Systematic Mapping Studies in Software

Engineering, Proc. of Intl Conference on Evaluation and Assessment in Soft. Eng., pp. 68 -77.

[22] Kitchenham, B. (2010) Whats up with Software Metrics? A Preliminary Mapping Study, J. Syst.Softw. 83 (1), pp. 37-51.

[23] Easterbrook, S., Singer, J., Storey, M., Damian, D. (2008) Guide to Advanced Empirical Software

Engineering, Selecting Empirical Methods for Soft. Engineering Research, Springer.

[24] Bratthall, L., Jrgensen, M. (2001) Can you Trust a Single Data Source Exploratory Software

Engineering Case Study?, Empirical Softw. Eng. 7 (1), pp. 705 -754.

[25] Shull, F., Basili, V., Carver, J., Maldonado, J. (2002) Replicating Software Engineering

Experiments: Addressing the Tacit Knowledge Problem, Proc. of Intl Symposium on Empirical

Software Engineering, pp. 7-16.

[26] Classen, A., Heymans, P., Schobbens, P., Legay, A. (2010) Symbolic Model Checking of Software

Product Lines, Proc. of Intl Conference of Software Engineering, 321 -330.

[27] Seaman, C. B. (1999) Qualitative Methods in Empirical Studies of Software Engineering, IEEE

Trans. Softw. Eng. 25 (4), pp. 557-572.

[28] Kitchenham, B., Pickard, L., Pfleeger, S. L. (1995) Case Studies for Method and Tool Evaluation,

IEEE Software 12 (4), pp. 52-62.[29] Soffer, P., Reinhartz-Berger, I., Sturm, A. (2007) Facilitating Reuse by Specialization of Reference

Models for Business Process Design, Proc. of Workshop on Business Process Modeling,

Development, and Support, pp. 339-347.

[30] Fantinato, M., Toledo, M. B. F., Thom, L. H., Gimenes, I. M. S., Rocha, R. S., Garcia, D. Z. G.

(2012) A Survey on Reuse in the Business Process Management Domain. Intl Journal of Business

Process Integration and Management 6, pp. 52-76.

[31] Rosenmller, M., Siegmund, N., Apel, S., Saak, G. (2011) Flexible Feature Binding in Software

Product Lines, Automated Software Engineering 18 (2), pp. 163 -197.

[32] Patil, M.V., Yogi, A.M.N. (2011) Importance of Data Collection and Validation for Systematic

Software Development Process. Int'l Journal of Computer Science & Inf. Technology, 3 (2).

[33] Aly, A.A., Deris, S.B., Zaki, N. (2011) Research Review for Digital Image Segmentation

Techniques. Int'l Journal of Computer Science & Inf. Technology 3 (5).

[34] Kara, M. (2012) Review on Common Criteria as a Secure Software Development Model. Int'l

Journal of Computer Science & Inf. Technology, 4 (2).


17/21


17

Appendix A Primary Studies

ID Title Author(s) Venue

P1A benchmarking framework formethods to design flexiblebusiness processes

F. Daoudi and S.Nurcan

Software Process: Improvement andPractice, 2007

P2A Conceptual Framework forIntention Driven FlexibleWorkflow Modeling

S. Nurcan

or s op on us ness o e ng,Development and Support - IntlConference on AdvancedInformation Systems Engineering,2004

P3 A Configurable ReferenceModelling LanguageM. Rosemann and W.M. P. van der Aalst Information Systems, 2007

P4A flexible, Object-centricApproach for Business ProcessModelling

. e ng, .Dumas, A. H.Hofstede and A.Iordachescu

Service Oriented Computing andApplications , 2010

P5A Role-Based Approach forModeling Flexible Business

Processes

O. Saidani and S.Nurcan

nt on erence on A vanceInformation Systems Engineering,

2006

P6A Survey on the FlexibilityRequirements Related to BusinessProcesses and Modeling Artifacts

S. Nurcan Hawaii Intl Conference on SystemSciences Engineering, 2008

P7 Achieving Business ProcessFlexibility with Business Rules

. van n oven,M. E. Iacoband M. L. Ponisio

Intl IEEE Enterprise DistributedObject Computing Conference, 2008

P8Achieving flexibility in BusinessProcess Modeling Using anAlgebraic Language

L. Xiao, B.H.Y. Kooand L. Zheng

Intl Conference on Model-BasedSystems Engineering, 2009

P9 Alternative Approaches forWorkflow SimilarityA. Wombacher andC. Li

IEEE Intl Conference on ServicesComputing, 2010

P10An Automation Support forCreating Configurable ProcessModels

W. Derguech and S.Bhiri

Intl Conference on WebInformation System Engineering,2011

P11 Analyzing the Scope of a Changein a Business Process Model P. Soffer

or s op on us ness o e ng,

Development and Support -IntlConference on AdvancedInformation Systems Engineering,2004

P12 Business Process Families UsingModel-Driven TechniquesV. Kulkarni and S.Barat

or s op on euse n us nessProcess Management - BusinessProcess Management Conference,2010

P13Business Process Flexibility:Weicks Organizational Theory tothe Rescue

G. Regev and A.Wegmann

us ness o e ng, eve opmentand Support Workshop IntlConference on AdvancedInformation Systems Engineering,2006

P14 Business Process Lines to Dealwith the VariabilityC. Rolland and S.Nurcan

Hawaii Intl Conference on SystemSciences, 2010

P15

Business Process Lines to developService-Oriented Architectures

through the Software ProductLines paradigm

. o o , .Caivano, D.Castelluccia, F. M.

Maggi and G.Visaggio

Software Product Line Conference,

2008

P16Business Process Model Merging:An Approach to Business ProcessConsolidation

. a osa, .Dumas, R. Kaarikand R. Dijkman

QUT ePrints Technical Report38241, Queensland University ofTechnology, 2008

P17 Business Process Modeling Awareto the Environment

N. Boffoli, D.Castelluccia, F. M.Maggi and R. Rutilo

nt on erence on va uat on oNovel Approaches to Software

Engineering, 2008

P18 Business Process Modelling andFlexibility S. NurcanIntl Conference on Interoperabilityfor Enterprise Software andApplications, 2007


18/21


18

P19Capturing Variability in BusinessProcess Models - the ProvopApproach

A. Hallerbach, T.Bauer and M. Reichert

Journal of Software Maintenanceand Evolution: Research andPractice, 2010

P20 Compliant and Flexible BusinessProcesses with Business RulesS. Goedertier and J.Vanthienen

CEUR Workshop and Practice,2006

P21

Conceptual Method for Flexible

Business Process

A. Bentellis and Z.

Boufada

World Academy of Science,

Engineering and Technology, 2008

P22 Configurable multi-perspectivebusiness process models

. a osa, .Dumas, A.H.M. terHofstede, and J.Mendling

Information Systems, 2011

P23 Configurable Process Models F. Gottschalk PhD thesis, Technische UniversiteitEindhoven, 2009

P24 Configurable Process Models: AFoundational ApproachF. Gottschalk and M.H. Jansen-vullers

Reference Modelling Conference,2006

P25 Configurable Process Models as aBasis for Reference Modeling

. . . van erAalst, A. DreilingM. Rosemann and

M.H. Jansen-Vullers

Workshop on Business ProcessReference Model - Business ProcessManagement Conference, 2005

P26

on gurat ve rocess o e ng:Outlining an Approach to

Increased Business Process ModelUsability Jrg Becker

Information Resources Management

Association Intl Conference, 2004

P27Configuring the Variability ofBusiness Process Models UsingNon-Functional Requirements

E. Santos, J. Pimentel,J. Castro, J. Snchezand O. Pastor

Exploring Modelling Methods forSystems Analysis and Design -Conference on AdvancedInformation Systems Engineering,2010

P28Context-aware Process DesignExploring the Extrinsic Driversfor Process Flexibility

M. Rosemann and J.Recker

Intl Conference on AdvancedInformation Systems Engineering,2006

P29 Correctness of Business ProcessModels with Roles and Objects

J. Mendling, M. LaRosa, A.H.M terHofstede

QUT ePrints Technical Report, 2008

P30 Defining Adaptation Constraintsfor Business Process Variants

R. Lu, S. Sadiq, G.Governatori and X.Yang

Intl Conference on BusinessInformation Systems, 2009

P31Defining Business ProcessFlexibility with the Help of

Invariants

G. Regev, I. Bider and

A. Wegmann

Software Process: Improvement and

Practice, 2007

P32 Defining Requirements forBusiness Process FlexibilityK. Kumar and M. M.Narasipuram

or s op on us ness rocessModelling, Development, andSupport - Conference on AdvancedInformation Systems Engineering,2006

P33 Diagnosing Differences betweenBusiness Process Models R. M. DijkmanIntl Conference on Business ProcessManagement, 2008

P34Ensuring Correctness DuringProcess Configuration via PartnerSynthesis

W.M.P. van derAalst, N. Lohmannand M. La Rosa

BPM Center Report BPM, 2011

P35Exploring the Dimensions ofVariability - a RequirementsEngineering Perspective

. . . . as os, .Jiang, A. Lapouchnian,Y. Wang, Y. Yu andJ.C.S.d.P. Leite, J.Mylopoulos

Intl Workshop on VariabilityModelling of Software-intensiveSystems, 2007

P36

xten ng a us ness rocess

Modeling Tool with ProcessConfiguration Facilities: TheProvop Demonstrator

. e c ert, .

Rechtenbach, A.Hallerbach, and T.Bauer

Intl Conference on Business ProcessManagement, 2009

P37 Extending the Adaptability ofReference Models

I. Reinhartz-Berger,P. Soffer and A.Sturm

IEEE Transactions on Systems, Manand, Cybernetics, Part A: Systemsand Humans, 2010

P38Fast Business Process SimilaritySearch with Feature-basedSimilarity Estimation

Z. Yan, R. Dijkmanand P. Grefen

Intl Conference on On the Move toMeaningful Internet Systems, 2010


19/21


19

P39 Flexible Process ModelingThrough Decision Tables

N. Boffoli, D.Castelluccia, F. M.Maggi, and R. Rutilo

IASTED Intl Conference onSoftware Engineering, 2008

P40 From Feature Models to BusinessProcesses TablesI. Montero, J. Penaand A. Ruiz-ortes

IEEE Intl Conference on ServicesComputing, 2008

P41

Guaranteeing Soundness of

Configurable Process Variants inProvop

A. Hallerbach, T.

Bauer and M.Reichert

IEEE Conference on Commerce and

Enterprise Computing, 2009

P42 How To Detect Semantic BusinessProcess Model Variants?A. Koschmider and A.Oberweis

ACM symposium on Appliedcomputing SAC, 2007

P43Improving Business ProcessModels with Reference Models inBusiness-Driven Development

J. M. Kster, J. Koehlerand K. Ryndina

Intl Workshop on Business ProcessDesign - Business ProcessManagement Conference, 2006

P44 Incremental Workflow Mining forProcess FlexibilityE. Kindler, V. Rubinand W. Schfer

Workshop on Business ProcessModelling, Development, andSupport - Conference on AdvancedInformation Systems Engineering,2006

P45 Issues in Modeling ProcessVariants with Provop

A. Hallerbach, T.Bauer and M.Reichert

Intl Workshop on Business ProcessDesign - Business ProcessManagement Conference, 2008

P46Managing Business ProcessFlexibility and Reuse Through

Business Process Lines

N. Boffoli, M.Cimitile and F. Maria

Maggi

Conference on Software and DataTechnologies 2009

P47 Managing Process Variants in theProcess Life Cycle

A. Hallerbach, T.Bauer and M.Reichert

Technical Report. University ofTwente, Enschede, The Netherlands,2007

P48

Managing SOA System Variationthrough Business Process Linesand Process OrientedDevelopment

. o o , .Cimitile, F. M.Maggi,and G. Visaggio

Workshop on Service-OrientedArchitectures and Software ProductLines, 2009

P49Managing Variability in BusinessProcesses: an Aspect-orientedApproach

. ac a o, .Bonifcio, V. Alves, L.Turnes, and G.Machado

Intl Workshop on Early Aspects,2011

P50Managing Variability in Process-Aware Information Systems - PhDThesis

M. La RosaQueensland University ofTechnology, Brisbane, Australia,2009

P51Managing Variability in Workflowwith Feature Model CompositionOperators

M. Acher, P. Collet, P.Lahire and R. France

Intl Conference on SoftwareComposition, 2010

P52 Measuring Similarity betweenBusiness Process Models

B. Dongen, R.Dijkman and J.Mendling


P53Measuring Similarity betweenSemantic Business ProcessModels

M. Ehrig, A.Koschmider and A.Oberweis

Asia-Pacific Conference onConceptual Modelling, 2007

P54 Mining business process variants:Challenges, scenarios, algorithmsC. Li, M. Reichert andA. Wombacher

Data & Knowledge Engineering,2011

P55 Mining Reference Process Modelsand Their Configurations

F. Gottschalk, Wil M.Aalst, andM. H. Jansen-Vullers

OTM Confederated Intl Workshopsand Posters on On the Move toMeaningful Internet Systems, 2008

P56 Modeling Variability in BusinessProcess Models Using UMLM. Razavian and R.Khosravi

Intl Conference on InformationTechnology: New Generations,2008

P57 Modelling Business Process

Variability - Masters thesisM. Vervuurt University of Twente, 2007

P58 On Managing Business ProcessesVariantsR. Lu, S. Sadiq andG. Governatori

Data & Knowledge Engineering,2009

P59 On the Notion of Flexibility inBusiness Processes P. Soffer

Workshop on Business ProcessModeling, Design and Support - IntlConference on AdvancedInformation Systems Engineering,2005


20/21


20

P60

n t e yntax o e erenceModel Configuration -Transforming the C-EPC intoLawful EPC Models

. ec er, .Rosemann, W. M. P.van der Aalst and J.Mendling

or s op on us ness rocessReference Models - BusinessProcess Management Conference,2005

P61 Partial Process Models to ManageBusiness Process Variants

. asca au, A. Awa ,S. Sakr andM. Weske

Intl Journal of Business ProcessIntegration and Management, 2011

P62 Preserving Correctness DuringBusiness Process ModelConfiguration

. . . van er

Aalst, M. Dumas, F.Gottschalk and A. H.M. ter Hofstede, M. LaRosa, J. Mendling

Formal Aspects of Computing, 2010

P63 Process Family EngineeringModeling Variant-rich Processes

. ayer, . u , .Giese, T. Lehner,A. Ocampo, F.

Puhlmann, E. Richter,

A. Schnieders, J.

Weiland and M.

Weske

PESOA (Process FamilyEngineering in Service-OrientedApplications) - Report TR, 2005

P64 Process Flexibility PatternsN. Mulyar, W.M.P.van der Aalst and N.Russell

Technical report, BETA WorkingPaper Series, 2008

P65 Process Flexibility: a Survey of

Contemporary Approaches

M. H. Schonenberg,R. S. Mans, N. C.

Russell and N. A.Mulyar

t nt or s op on nterpr seand Organizational Modeling AndSimulation - Intl Conference on

Advanced Information SystemsEngineering, 2008

P66 Process Flexibility: Design,Evaluation, and Applications

Mabel C. Chou,Chung-Piaw Teo andHuan Zheng

Flexible Services and ManufacturingJournal, 2008

P67 Process Mining by MeasuringProcess Block SimilarityJ. Bae, J. Caverlee, L.Liu and H. Yan

Workshop on Business ProcessIntelligence - Business ProcessManagement Conference, 2006

P68Questionnaire-based VariabilityModeling for SystemConfiguration

. a osa, . .P. van der Aalst, M.Dumas and A. H. M.ter Hofstede

Information and System Modeling,2009

P69Reuse-Oriented Business ProcessModelling Based on aHierarchical Structure

W. Derguech, S. Bhiri

or s op on euse n us nessProcess Management - BusinessProcess Management Conference,2010

P70 Similarity of Business Process

Models: Metrics and Evaluation

. man, .Dumas, B. Van

Dongen, R. Krik and J.Mendling

Information Systems, 2011

P71 Similarity Search of BusinessProcess Models

. umas, . arca-bauelos, R. DijkmanR. Krik and J.Mendling

IEEE Data Engineering Bulletin,2009

P72Structural and BehaviouralCommonalities of ProcessVariants

. e c , .Weske, R. Krik andJ. Mendling

Zentral-europischer Workshop berServices und ihre Komposition,2010

P73

Support for Business ProcessFlexibility in ServiceCompositions: An EvaluativeSurvey

M. Kapuruge, J. Hanand A. Colman J.Mendling

Australian Software EngineeringConference, 2010

P74 Supporting Planned and Ad-HocChanges of Business Processes P. Rittgen

or s op on us ness rocessModelling, Development, andSupport - Intl Conference onAdvanced Information SystemsEngineering, 2006

P75 Taxonomy of Flexibility inBusiness ProcessesG. Regev, P. Sofferand R. Schmid

Workshop on Business Process

Modelling, Development, andSupport - Intl Conference onAdvanced Information SystemsEngineering, 2006

P76Towards a Classification andLifecycle of Business ProcessChange

K. Ploesser, J. Reckerand M. Rosemann

or s op on us ness rocessModeling, Development andSupport - Conference on AdvancedInformation Systems Engineering,2008

P77

owar s A apta ty an ontrofor Knowledge-Intensive BusinessProcesses: DeclarativeConfigurable Process

I. Rychkova and S.Nurcan

Hawaii Intl Conference on SystemSciences, 2011


21/21


21

pec cat ons

P78Towards E-government byBusiness Process Change - AMethodology for Public Sector

M. I. Stemberger andJ. Jaklic

Intl Journal of InformationManagement, 2007

P79 Validation of Families of BusinessProcesses

G. Grner, C. Wende,M. Boskovic, F. S.Parreiras, T. Walter,F. Heidenreich, D.Gasevic and S. Staab


P80 Variability Mechanisms forProcess Models

F. Puhlmann, A.Schnieders, J. Weilandand M. Weske

PESOA (Process FamilyEngineering in Service- OrientedApplications) - Report TR, 2005

AUTHORS

George Valena received his masters degree in Computer Science in 2012 from the

Federal University of Pernambuco (UFPE)/Brazil, addressing business process variability

topic. He is currently a PhD student at UFPE, studying Software Ecosystems field. He also

has experience in Business Process Management, Requirements Engineering and Software

Quality topics.

Carina Alves is adjunct professor of the Informatis Center (CIn) of UFPE. She received the

PhD degree in Computer Science in 2005 from the University College London/UK. She is

currently head of the Computer Systems Department at CIn/UFPE. She studies the fields of

Requirements Engineering, Business Process Management and Software Ecosystems.

Vander Alves is adjunct professor of the Computer Science Department of the University of

Brasilia/Brazil. He received his PhD degree in Computer Science in 2007 at UFPE, with

postdoctoral research at Lancaster University/UK in 2008 and at Fraunhofer

Institute/Germany in 2009. He has experience in Software Product Line, Formal Methods

and Ambient Assisted Living topics.

Nan Niu is assistant professor of the Department of Computer Science and Engineering of

the Mississippi State University/USA. He received his PhD degree in Computer Science in2009 at University of Toronto/Canada. He investigates the fields of Requirements

Engineering, Program Comprehension and Software Reuse.

Documents

A Systematic Mapping Study on Business Process Variability